From 81a0624d7613dbd5a4fc69cde95314e7db47bd43 Mon Sep 17 00:00:00 2001 From: Alois Wohlschlager Date: Mon, 15 Jul 2024 18:00:29 +0200 Subject: [PATCH] libexpr/print: pretty-print idempotently MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When pretty-printing is enabled, previously an unforced thunk would trigger indentation, even when it subsequently does not evaluate to a nested structure. The resulting output looked inconsistent, and furthermore pretty-printing was not idempotent (since pretty-printing the same value again, which is now fully evaluated, will not trigger indentation). When strict evaluation is enabled, force the item before inspecting its type, so that it is properly known whether it contains a nested structure. Furthermore, there is no need to cause indentation for unforced thunks, since the very next operation will be printing them as `«thunk»`. This is mostly a port of https://github.com/NixOS/nix/pull/11100 , but we only force the item when it's going to be forced anyway due to strict pretty-printing, and a new test was written since the REPL testing framework in Lix is different. Co-Authored-By: Robert Hensing Change-Id: Ib7560fe531d09e05ca6b2037a523fe21a26d9d58 --- src/libexpr/print.cc | 14 ++++++++++++-- .../repl_characterization/data/idempotent.test | 13 +++++++++++++ .../repl_characterization/repl_characterization.cc | 1 + 3 files changed, 26 insertions(+), 2 deletions(-) create mode 100644 tests/functional/repl_characterization/data/idempotent.test diff --git a/src/libexpr/print.cc b/src/libexpr/print.cc index 87db004b2..6eb321c43 100644 --- a/src/libexpr/print.cc +++ b/src/libexpr/print.cc @@ -264,10 +264,15 @@ private: return true; } + if (options.force) { + // The item is going to be forced during printing anyway, but we need its type now. + state.forceValue(*item, item->determinePos(noPos)); + } + // Pretty-print single-item attrsets only if they contain nested // structures. auto itemType = item->type(); - return itemType == nList || itemType == nAttrs || itemType == nThunk; + return itemType == nList || itemType == nAttrs; } void printAttrs(Value & v, size_t depth) @@ -335,10 +340,15 @@ private: return true; } + if (options.force) { + // The item is going to be forced during printing anyway, but we need its type now. + state.forceValue(*item, item->determinePos(noPos)); + } + // Pretty-print single-item lists only if they contain nested // structures. auto itemType = item->type(); - return itemType == nList || itemType == nAttrs || itemType == nThunk; + return itemType == nList || itemType == nAttrs; } void printList(Value & v, size_t depth) diff --git a/tests/functional/repl_characterization/data/idempotent.test b/tests/functional/repl_characterization/data/idempotent.test new file mode 100644 index 000000000..4ab087d45 --- /dev/null +++ b/tests/functional/repl_characterization/data/idempotent.test @@ -0,0 +1,13 @@ +A previously unforced thunk in an attribute set does not lead to indentation when it won't evaluate to a nested structure: + nix-repl> :p let x = 1 + 2; in [ { inherit x; } { inherit x; } ] + [ + { x = 3; } + { x = 3; } + ] + +Same for a list: + nix-repl> :p let x = 1 + 2; in [ [ x ] [ x ] ] + [ + [ 3 ] + [ 3 ] + ] diff --git a/tests/functional/repl_characterization/repl_characterization.cc b/tests/functional/repl_characterization/repl_characterization.cc index c91d6c1e3..f2cf0ca3d 100644 --- a/tests/functional/repl_characterization/repl_characterization.cc +++ b/tests/functional/repl_characterization/repl_characterization.cc @@ -185,5 +185,6 @@ REPL_TEST(repl_overlays_error); REPL_TEST(repl_printing); REPL_TEST(stack_vars); REPL_TEST(errors); +REPL_TEST(idempotent); }; // namespace nix